diff --git a/Demo/DEMO_generate_random_mask.py b/Demo/DEMO_generate_random_mask.py
index 9181e23..ee4e0e2 100644
--- a/Demo/DEMO_generate_random_mask.py
+++ b/Demo/DEMO_generate_random_mask.py
@@ -1,28 +1,25 @@
 import cv2
 import numpy as np
-# https://github.com/Yonv1943/Python/tree/master/Demo
-# https://www.zhihu.com/search?type=content&q=EdgeConnect
-# DEMO_generate_random_mask.py
 
+# Source: LearningToPaint - hzwer
+# Modify: Yonv1943 Github
+# DEMO_generate_random_mask.py
 
-def brush__bezier_curve(thetas, width1=128):
-    # Source: LearningToPaint - hzwer
-    # Modify: Yonv1943 Github
+def brush_bezier_curve(thetas, width1=128):
     # paras = np.random.rand(128, 10)
     width2 = width1 * 2
     thetas[:, 0:4] *= width2
     thetas[:, 4:6] = thetas[:, 0:2] + (thetas[:, 2:4] - thetas[:, 0:2]) * thetas[:, 4:6]
-    thetas[:, 6:8] = thetas[:, 6:8] * (width2 // 8) + 2  # add 2 to ensure the strokes are not to thin.
-    thetas[:, 8:10] *= 255  # max_uint8 == 255
+    thetas[:, 6:8] = thetas[:, 6:8] * (width2 // 8) + 2
+    thetas[:, 8:10] *= 255
 
     res = np.empty((thetas.shape[0], width1, width1), dtype=np.uint8)
     for idx, para in enumerate(thetas):
         x0, y0, x2, y2, x1, y1, z0, z2, w0, w2 = para
-        # staring point, ending point, middle point, thickness, color
 
         canvas = np.zeros([width2, width2], dtype=np.uint8)
         gap = width1 // 2
-        for p in range(gap):  # bezier curve
+        for p in range(gap):
             p /= gap
             q = 1 - p
 
@@ -35,33 +32,22 @@ def brush__bezier_curve(thetas, width1=128):
             z = int(p * z2 + q * z0)
             w = int(p * w2 + q * w0)
 
-            cv2.circle(canvas, (y, x), z, w, -1)  # img, center, radius, color,
+            cv2.circle(canvas, (y, x), z, w, -1)
         res[idx] = cv2.resize(canvas, dsize=(width1, width1))
 
-    # res = res[:, np.newaxis, :, :]
     return res
 
-
 def generate_random_mask(thetas, width2=128):
     # paras = np.random.rand(128, 8)
-
-    # thetas[:, 0:2]: starting point (x0, y0)
-    # thetas[:, 2:4]: ending point (x2, y2)
     thetas[:, 0:4] *= width2
-
-    # thetas[:, 4:6]: middle point, it stay between starting points and ending points
     thetas[:, 4:6] = thetas[:, 0:2] + (thetas[:, 2:4] - thetas[:, 0:2]) * thetas[:, 4:6]
-
-    # thetas[:, 6:8]: the thickness of the strokes(mask)
-    # add 2 to ensure the strokes are not to thin.
     thetas[:, 6:8] = thetas[:, 6:8] * (width2 // 8) + 2
 
-    # draw the random strokes(mask)
     canvas = np.ones([width2, width2], dtype=np.uint8)
     gap = width2 // 2
     for idx, para in enumerate(thetas):
         x0, y0, x2, y2, x1, y1, z0, z2 = para
-        for p in range(gap):  # bezier curve
+        for p in range(gap):
             p /= gap
             q = 1 - p
 
@@ -73,21 +59,19 @@ def generate_random_mask(thetas, width2=128):
             y = int(pp * y2 + pq2 * y1 + qq * y0)
             z = int(p * z2 + q * z0)
 
-            cv2.circle(canvas, (y, x), z, 0, -1)  # img, center, radius, color,
+            cv2.circle(canvas, (y, x), z, 0, -1)
 
     return canvas
 
-
 if __name__ == '__main__':
     WIDTH = 128
     Mask = generate_random_mask(np.random.rand(2, 8), WIDTH)
     Mask = Mask[:, :, np.newaxis]
 
-    # Img = np.random.randint(0, 255, (WIDTH, WIDTH, 3), dtype=np.uint8)
     Img = cv2.imread('image/lisa.png')
     cv2.imshow('Random Mask', Img * Mask)
     cv2.waitKey(12340)
 
-    for image in brush__bezier_curve(np.random.rand(8, 10), WIDTH):
+    for image in brush_bezier_curve(np.random.rand(8, 10), WIDTH):
         cv2.imshow('Random Mask', image)
         cv2.waitKey(234)